Your ERP and your spend tools show what you paid. Cotiss shows what you agreed, what it means, and what to do about it.
The Platform
Cotiss is built for one job: turning the unstructured material your commercial function generates every day into a structured, queryable, audit-grade record.
01
Reads what no other system can: the sourcing events, vendor conversations, contract negotiations, and exceptions that shape a decision, captured as what was decided and why.
Emails, SharePoint, meetings, contracts, and messages, all under role-based scope.
02
Reasons over it through your own rules, building a connected model of your commercial reality: categories, vendors, contracts, clauses, policies, regional rules, and delegations.
Your existing systems weren't built to hold this, and nor were most enterprise data lakes.
03
Holds the reasoning across people, time, and turnover. Decisions from 2022 stay queryable in 2030.
The longer it runs, the more of your reasoning it holds. Answers that lived in one person's head outlast them.
04
Runs the work off that memory, acting when you say go, with a named person accountable for anything it does.
Works alongside the systems you already run.
Read
The ERP and the contract lifecycle system hold what got typed in after the fact. Most of what actually happens with a vendor happens before that: in the email where a price was negotiated down, in the call where a delivery date slipped, in the meeting notes where someone agreed to a scope change nobody logged anywhere else.
Reason
Write your rules, thresholds, and playbooks in plain English. Cotiss reads every decision against them and acts on the ones that don't fit.
Most finance and procurement teams have hundreds of pages of policy, dozens of strategies, and a handful of unwritten rules everyone just knows. None of it reaches the systems where the decisions actually get made, so Cotiss turns that knowledge into the rules the system runs on.
No code, no config files, no consultants. Plain English, written the way you'd brief a new hire, and updated whenever your rules change.
Book a demo →01
Never let a vendor raise prices past the cap you agreed to. Catch it in the email, before it hits an invoice.
What Cotiss does
Cotiss checks every increase against the escalator clause the moment it lands, and drafts the response for you to send.
Remember
Cotiss arrives with a working ontology and then learns yours: your vendor naming conventions, your category terminology, the policy exceptions everyone already knows about, and the way your business actually talks about the people it pays.
That's what makes the reasoning accurate, because the structure is drawn from your own documents and keeps fitting as you grow.
Step 01
Vendors, categories, clauses, commitments, and exceptions are first-class concepts, so the system understands commercial work before it ever sees yours.
Step 02
We map your terminology, structure, and exceptions during implementation, so the reasoning fits your business rather than a template shared across every customer.
Step 03
Vendors get renamed and categories get restructured every time a team reorganizes, so the structure follows your business instead of dating from the day someone first set it up.
Run
You describe an automation the way you'd say it out loud: when this happens, check these things, then do that. A price increase arrives by email, Cotiss checks it against the contract it falls under, and sends it to the person who owns that vendor with the clause already attached.
The trigger is the part that needs comprehension. A generic automation tool waits for a field to change in a database, which means someone has already read the email and typed the fact in. Cotiss starts while the fact is still prose, the moment it understands what an email or a PDF says.
Cotiss tells you what it found and stops there. Nothing moves without you.
Cotiss drafts the action and holds it for a decision. You send it, change it, or drop it.
Cotiss does the work and leaves the record, and the reasoning, for you to check afterwards.
Cotiss handles it inside the boundaries you set, and escalates anything that falls outside them.
You can dry run an automation against your own recent history before it goes live. A named person stays accountable for whatever it does, and every action keeps its evidence, so a cautious team starts at the first rung and moves up at its own pace.
Your close tool, your AP platform, and your procurement suite keep running the work. Cotiss acts on the record it already holds.
Transparency
Every figure traces back to the clause on the page, the sentence in the email, or the second in the call it came from.
Every answer arrives with its evidence attached, so you never have to decide whether to believe it.
A confidence score isn't an answer and a cited document is. The product was built around that from the first day.
Drill Down
Drafted against your escalator rule and sent by the category manager.
Decided 14 Mar 11:40
Effective 1 April, list pricing will increase by 6% across all SKUs.
Received 14 Mar 09:04
shall not exceed three percent (3%) in any twelve-month period
Read 14 Mar 09:12 · page 4
Security
From the blog
The difference is what Cotiss is built for: the reasoning layer your systems of record and agent workflows miss.
Read the full post →Structured-data context
Most tools enrich what already lives in ERP, P2P, and contract systems. Useful, but it still misses the reasoning behind the data.
AI agents over the data
Agents can draft, recommend, and automate, but they need a durable record of why decisions were made to act with real understanding.
Unstructured-reasoning context
Cotiss captures the why behind commercial decisions from emails, documents, meetings, and conversations, then makes it queryable.
Where we start
Runs alongside your existing stack · Usage-based pricing · From data to decision in three weeks